广告合作
MCP.so
登录

Firebase MCP Server

@gannonh

代码仓库

基本信息

分类

云与基础设施

许可证

MIT

传输方式

stdio

作者

gannonh

关于 Firebase MCP Server

🔥 Model Context Protocol (MCP) server for Firebase.

#firebase#modelcontextprotocol

Firebase MCP Server 配置

使用下面的配置,将此服务器添加到你的 MCP 客户端。

{
  "mcpServers": {
    "firebase-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "@gannonh/firebase-mcp"
      ],
      "env": {
        "SERVICE_ACCOUNT_KEY_PATH": "/absolute/path/to/serviceAccountKey.json",
        "FIREBASE_STORAGE_BUCKET": "your-project-id.firebasestorage.app"
      }
    }
  }
}

Firebase MCP Server 工具

工具列表暂未提供

当前收录暂时没有可展示的工具列表，不代表服务没有工具。请查阅项目文档了解支持的能力。

点击上方按钮,在临时沙箱中运行此服务器并实时获取工具列表。

概览

What is Firebase MCP?

Firebase MCP enables AI assistants to work directly with Firebase services, including Firestore document operations, Storage file management with upload capabilities, and Authentication user management and verification. It works with MCP client applications such as Claude Desktop, Augment Code, VS Code, and Cursor.

How to use Firebase MCP?

Add the server configuration to your MCP client’s settings file (e.g., claude_desktop_config.json). Provide the required environment variable SERVICE_ACCOUNT_KEY_PATH pointing to your Firebase service account key JSON, and optionally set FIREBASE_STORAGE_BUCKET. The server can be run via npx (recommended) or from a local installation, and supports both stdio and HTTP transports.

Key features of Firebase MCP

  • Firestore documents: add, list, get, update, delete
  • Firestore collections: list root collections and query collection groups
  • Storage files: list, get metadata, upload from content or URL
  • Authentication: get user by ID or email
  • HTTP transport with session management for multiple clients
  • Optional file logging for debugging

Use cases of Firebase MCP

  • Query and manage Firestore data through natural language conversation
  • Upload files to Firebase Storage and retrieve download URLs
  • Look up user details in Firebase Authentication
  • Perform CRUD operations on documents without writing code
  • Debug Firebase queries with file logging and MCP Inspector

常见问题

1

What prerequisites are needed to run Firebase MCP?

You need a Firebase project with a service account key (JSON), and a Node.js environment.

2

Does the `firestore_list_collections` tool have a known issue?

Yes. It may show a Zod validation error in client logs, but the query still works correctly; this is a log-level bug in the MCP SDK that does not affect functionality.

3

How do I fix a “Storage Bucket Not Found” error?

Verify your bucket name in the Firebase Console and set the correct value in the `FIREBASE_STORAGE_BUCKET` environment variable.

4

What should I do if I get a “composite index required” error?

Follow the URL provided in the error message to create the required composite index in the Firebase Console, then retry the query after the index is ready.

5

How can I enable debugging for Firebase MCP?

Set `DEBUG_LOG_FILE=true` in the environment to log to `~/.firebase-mcp/debug.log`, or specify a custom path for file logging. You can also use the MCP Inspector for interactive debugging.

评论

云与基础设施 分类下的更多 MCP 服务器

monadomain-mcp

themonagroup

Buy .vn and 371 international domains in VND inside Claude Code, Codex or Cursor. Search and reserve as a guest, pay by VietQR, sign in with MONA Pass once to claim; the agent then points DNS and SSL at your app. Alias o

Tessryx

Build and run websites, APIs, automations, and admin tools from chat. Your agent creates the data, logic, pages, and scheduled jobs; Tessryx hosts them on your own domain. No code generated, nothing to deploy.

Azure DevOps MCP Server

Tiberriver256

An MCP server for Azure DevOps

377

OutageDeck

outagedeck

Official OutageDeck MCP server for vendor-published cloud and SaaS status, incident timelines, and observed uptime history.

Kilo IoT Platform

Kilo IoT

Your AI can finally reach the physical world. Kilo is an AI-first IoT platform for commercial and industrial operations, and this connector hands a live deployment to the AI client you already use. Not an export. Not a

广告合作